The goal of this research is to identify reasons for variability in the interpretation of mammograms. Though previous studies have shown marked interpretive variation, they did not explain why it occurs and they used test sets that do not necessarily reflect what occurs in day-to-day community practices. This community-based, multicenter observational study will utilize a unique collaboration among three geographically distinct breast cancer surveillance programs in the states of Washington, New Hampshire, and Colorado. This collaboration will allow the investigators to accumulate breast cancer outcome and interpretive data on more than 500,000 mammograms from 91 facilities and 279 radiologists. The investigators will evaluate potential factors influencing the accuracy and recall rates of mammography using a structured conceptual framework that separates characteristics of the radiologists from those of the facility and community environment. The investigators hypothesize that the fiscal, legal, and community practice environment and personal characteristics of the radiologist influence variability in the accuracy of mammography and the likelihood of having patients recalled for additional evaluation. Gaining a better understanding of how individual radiologists and their practice environment account for variation will help identify ways to improve mammography.